On Mon, 11 Jul 2011 14:56:26 +0100, Murray S. Kucherawy
<msk(_at_)cloudmark(_dot_)com> wrote:
-----Original Message-----
From: ietf-dkim-bounces(_at_)mipassoc(_dot_)org
[mailto:ietf-dkim-bounces(_at_)mipassoc(_dot_)org] On Behalf Of Charles
Lindsey
Sent: Monday, July 11, 2011 3:52 AM
"Agents that evaluate or apply DKIM output need to be aware that a
DKIM
signer can sign messages that are malformed (e.g., violate RFC5322),
or
become malformed in transit, or contain content that is not true or
valid. Such an action might constitute an attack against a receiver,
especially where additional credence is incorrectly given to a signed
message without evaluation of the signer. Moreover, an agent would be
incorrect to infer that all instances of a header field are signed
just
because one is. Agents will need to account for these issues when
deciding how to apply DKIM results to message, especially when
displaying them to users."
OK, there is much good stuff in that. In particular, it makes it clear
that Bad Stuff can originate from the signer as well as from
men-in-the-middle and replayers. But I am still concerned that multiple
occurrences of "singleton" headers fields are not explicitly mentioned,
even as just one possible example.
I still don't think that paragraph is what we really need, but I will
withold judgement on that until I see how it gets incorporated into the
other bits of text that are around.
Given that today's the deadline, we will have to go with something like
this or nothing at all (which in fact I would prefer because I think all
of this is adequately covered by existing text, and I believe consensus
and the AD concurs), so withhold judiciously.
Essentially, my concern is that an implementor reading this section should
be able to infer the nature of the particular attack I have described (the
one where the signer is the BadGuy himself using a throwaway domain),
including spotting how and why it worked and how to protect against it.
Having now read your paragraph in the context of the rest of the material
in that section, I think it just about passes that test, but only by the
thinnest of margins, so I will let it go.
But as a piece of technical writing that section is a total mess, talking
around the problem, and seemingly more concerned with enabling timid
implementors to pass the buck around amongst themselves that with
protecting the ultimate users.
I see Doug has written a detailed critique of it, and I fully endorse most
of what he has said.
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131
Web: http://www.cs.man.ac.uk/~chl
Email: chl(_at_)clerew(_dot_)man(_dot_)ac(_dot_)uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html